Skip to main content
KBY Technologies Logo
All Engineering Labs
SimulatorPrincipal

Incident Command Simulator

A timed decision simulator for practising command, containment and recovery during a multi-region service incident. It focuses on clear roles, reversible actions, evidence preservation and disciplined recovery gates.

Around 8 min Saved private report Principal level
Purpose and audience

Who this Lab is for

Designed for

  • Incident commanders and technical leads
  • Senior on-call engineers
  • Teams rehearsing major-incident procedures

Use it when

  • Preparing engineers for incident-command responsibility
  • Running individual practice between team game days
  • Reviewing command principles after an incident
How to use it

A complete run, step by step

1

Select the incident

Choose payment degradation, identity failure or queue saturation.

2

Make the first command decision

Establish control, assign roles and protect evidence while the situation is uncertain.

3

Choose containment

Limit blast radius with the most reversible action supported by current evidence.

4

Set a recovery gate

Require customer and system signals to stabilise before declaring resolution or resuming changes.

Input guide

What you will need

Prepare the following information before starting. Use measured evidence where possible; defaults are examples and should not be treated as recommendations.

Available scenarios

Payment degradation

Connection-pool exhaustion following a client rollout.

Identity outage

Token validation failures spread across regions.

Queue saturation

Retry amplification overwhelms downstream consumers.

Minute 2 — first command

select

Latency rises after a database client rollout.

Choices: Freeze changes, assign roles, preserve telemetry · Restart every database node · Wait for more customer reports

Minute 8 — containment

select

One region shows connection-pool exhaustion.

Choices: Rollback the client in the affected region · Fail all traffic into the degraded region · Double every pool limit

Minute 18 — recovery gate

select

Error rate falls but queues remain elevated.

Choices: Verify saturation, drain rate and customer errors · Declare resolved immediately · Resume deployments

Results and methodology

What the result tells you

Your report includes

  • A command-decision score
  • Feedback on containment and recovery discipline
  • A saved record across multiple incident variants

How it is determined

Decisions are weighted for command clarity, evidence preservation, reversibility, customer-risk reduction and verification before recovery. Fast but uncontrolled actions score poorly even when they could sometimes work.

Indicative rubric · low confidence · v2026.07.1

The score evaluates three command choices in a simplified scenario and cannot assess team communication or live telemetry.

Model assumptions

  • Reducing concurrent change is appropriate during uncertainty.
  • Targeted rollback is supported by the stated evidence.
  • Recovery requires customer and system signals to stabilise.
Worked example

Payment degradation after rollout

Situation

Latency rises after a database client release and one region shows connection-pool exhaustion.

Result

Freezing changes, assigning roles and rolling back only the affected region produces a controlled path to recovery with explicit verification gates.

Important limitations

Use the result with engineering judgement

  • The simulator cannot reproduce team communication or incomplete telemetry.
  • Real incidents require your approved command structure and escalation paths.
  • A high score does not certify someone to command an incident alone.
Frequently asked questions

Questions before you begin

Is this a game day?

It is individual decision practice. Combine it with facilitated team game days for communication, tooling and coordination practice.

Why favour reversible actions?

Early incident hypotheses are uncertain. Reversible containment reduces the chance of making impact worse while evidence develops.

When is an incident resolved?

When customer impact has ended, key system signals are stable and the recovery state has been verified—not simply when one metric improves.

Incident Command is under review

This legacy judgement-based Lab has been retired. Existing saved reports remain available, but new execution is disabled.

Open deterministic utilities